Digital twins: what are they, what are they for and what are the benefits and problems of digital twins?

How do NASA engineers ensure that they will be able to control any element of a special spacecraft when it is thousands of miles away? That they will know how it behaves when other temperatures and gravity forces come into play? Creating digital twins.

What are digital twins?

A digital twin is a digital representation of a physical object, process or service: from a jet engine or wind farm to buildings or entire cities. These virtual replicas are used to run simulations before changes are created and implemented on real objects, in order to collect data to predict how they will function.

In many occasions, physical objects are equipped with many sensors in charge of collecting data about their real-time status, working conditions or position. All this data must be analyzed and processed to recreate the digital model.

How it is generated

A digital twin is created in a computer program that uses real-world data to recreate simulations that can predict how a product or process will work. They are used to prevent failures in physical objects and to perform advanced analysis, monitoring and prediction functions.

Those responsible for creating a digital twin are usually experts in applied mathematics or data science. These professionals analyze the physics and operational data of an object and develop a mathematical model that simulates the original.

When creating these digital twins, it must be ensured that the virtual model can receive feedback from sensors that collect data from the real-world version. This allows the digital version to mimic and simulate what is happening with the original version in real time.

In this technical tutorial you can see how to create a digital twin of a car.

What you need to create them

To create a digital twin you need to collect a lot of data, both about the object and what is around it. With this information, computational models can be created that represent the behaviors or states of the physical object.

This data can be related to the life cycle of a product, its design specifications, its production processes, engineering information, production information (including materials, parts, methods and quality control)…

A digital twin can be as complex or as simple as needed. The amount of data you collect will also determine how accurately the digital model simulates the physical version.

What are they used for?

Once all the data has been collected, it is used to create analytical models to predict the effects and behavior of that object under possible changes.

These simulations are generated taking into account engineering, physics, chemistry, statistics, machine learning, artificial intelligence, business logic or objectives. These models can be displayed through 3D renderings and augmented reality modeling to aid human understanding of the findings.

The development and creation of a digital twin is broadly used for three main issues.

• Digital Twin Prototyping (DTP): before creating a final physical product, a digital one is made to see what it would actually look and behave like. 

• Digital Twin Instance (DTI): once a product has been manufactured, the digital twin is used to test different usage scenarios with the virtual rather than the real one.

• Digital Twin Aggregate (DTA): collects information from the previous case to determine the capabilities of a product, run forecasts and test operational parameters.

Through these three typical use cases companies can predict different outcomes based on variable data, which helps determine where things should go or how they work before they are physically implemented.

What are the benefits

The digital twin creates a simulation model that can be updated alongside or instead of the actual physical model. This allows companies to evaluate a fully computerized development cycle, from design to implementation and even decommissioning.

By mimicking physical assets, frameworks and operations to produce continuous data, a digital twin enables industry to anticipate downtime, react to changing circumstances, test design improvements and much more.

A digital twin can be used to save time and money whenever there is a need to test a product or process, whether in design, implementation, monitoring or improvement.

Obviously, these benefits of the digital twin are very case-dependent, but in large infrastructures it allows you to track and analyze existing products. By doing so, their maintenance and associated costs can be reduced, and potential failures can be foreseen and anticipated.

When used for prototyping prior to manufacturing, digital twins can reduce product defects, improving product planning and delivery, as well as product quality.

In which industries are they used

Digital twins are already being used in several industries, with different applications and purposes. For example, in factories they are used to make the production chain more agile and reduce potential errors.

The medical sector has benefited from the digital twin in areas such as organ donation, surgical training, or modeling the flow of people through hospitals and track where infections may exist and who may be at risk from contact.

In logistics, digital twins can be used in a wide variety of applications, such as managing container fleets, monitoring shipments, or designing large logistics systems. IoT sensors in individual containers, for example, display their location and monitor issues such as damage or contamination caused by their transport. This data flows to a digital twin of the container network, which uses machine learning to make the necessary improvements to ensure that containers are implemented as efficiently as possible.

DHL, for example, is one of the companies using this technology.

What problems do they have?

Building a digital twin is complex. In addition, there is no standardized platform for doing so, nor is it entirely clear what technology is needed to build and implement digital twins.

Commercial proposals for digital twins come from a few large companies. For example, GE, which developed digital twin technology internally as part of its jet engine manufacturing process, is now offering its expertise to other companies, as is Siemens, another industrial giant heavily involved in manufacturing.

But technology names such as IBM and Microsoft are also commercializing their digital twin proposition.

Who invented digital twins

The concept of digital twins is attributed to David Gelernter and his book “Mirror Worlds,” but it was Michael Grieves of the Florida Institute of Technology who applied it to the manufacturing concept in 2002.

However, it was NASA that first adopted the digital twin concept in 2010, for the purpose of maintaining and repairing systems when they were not around. The idea was used to create digital simulations of space capsules and spacecraft for testing. Today, NASA uses digital twins to develop new recommendations, roadmaps, and next-generation vehicles and aircraft.